Best printers for Home

Best Mixer Grinder

Best wired Earphones

Best 43 Inch TV in India

Best Wi Fi Routers

Best Home Theatre in India

Smart Watch under 5000

Best Laptops for Education

Best Laptop for Students

New York, In a significant experiment, a US medical school used Open AI's <keyword id="18153363" type="General" weightage="20" keywordseo="Chat-GPT-4" source="Orion">Chat-GPT 4</keyword> to see if it can make accurate diagnoses in challenging medical cases. Physician-researchers at Beth Israel Deaconess Medical Center (BIDMC) in Boston found that <keyword id="16832125" type="General" weightage="20" keywordseo="chat-gpt" source="Orion">Chat-GPT</keyword> 4 selected the correct diagnosis nearly 40 per cent of the time. Chat-GPT 4 also provided the correct diagnosis in its list of potential diagnoses in two-thirds of challenging cases, revealed the findings, published in JAMA. "Recent advances in artificial intelligence have led to generative AI models that are capable of detailed text-based responses that score highly in standardised medical examinations," said Adam Rodman, co-director of the Innovations in Media and Education Delivery (iMED) Initiative at BIDMC. "We wanted to know if such a generative model could 'think' like a doctor, so we asked one to solve standardised complex diagnostic cases used for educational purposes. It did really, really well," said Rodman, also an instructor in medicine at Harvard Medical School. To assess the chatbot's diagnostic skills, Rodman and colleagues used clinicopathological case conferences (CPCs), a series of complex and challenging patient cases including relevant clinical and laboratory data, imaging studies, and histopathological findings published in the New England Journal of Medicine for educational purposes. Evaluating 70 CPC cases, the artificial intelligence exactly matched the final CPC diagnosis in 27 (39 per cent) of cases. In 64 per cent of the cases, the final CPC diagnosis was included in the AI's differential -- a list of possible conditions that could account for a patient's symptoms, medical history, clinical findings and laboratory or imaging results. "While chatbots cannot replace the expertise and knowledge of a trained medical professional, generative AI is a promising potential adjunct to human cognition in diagnosis," said first author Zahir Kanjee, a hospitalist at BIDMC and assistant professor of medicine at Harvard Medical School. "It has the potential to help physicians make sense of complex medical data and broaden or refine our diagnostic thinking," he said. While the study adds to a growing body of literature demonstrating the promising capabilities of AI technology, more research is needed on its optimal uses, benefits and limits, importantly on privacy issues to understand how these new AI models might transform health care delivery. --IANS rvt/vd

ChatGPT was asked to think like a doctor and it did 'really well' say researchers

Americans are trapped in a loneliness epidemic. Across the country, people are having fewer social interactions, spending more time alone, and reporting <a href="https://www.americansurveycenter.org/research/the-state-of-american-friendship-change-challenges-and-loss/" target="_blank" rel="noopener">fewer close friends</a>. These trends aren't just a symptom of the COVID-19 pandemic — while the last few years may have accelerated the loneliness crisis, the shift toward a more solitary life <a href="https://www.vox.com/2020/5/11/21245087/america-loneliness-epidemic-coronavirus-pandemic-together" target="_blank" rel="noopener">has been happening for years</a>.A <a href="https://www.hhs.gov/sites/default/files/surgeon-general-social-connection-advisory.pdf" target="_blank" rel="noopener">new report</a> from the US surgeon general finds that social activities of all kinds have declined, and it compared the health impact of this increasing <a href="https://www.insider.com/us-surgeon-general-compares-loneliness-epidemic-to-smoking-2023-5">loneliness to smoking 12 cigarettes a day</a>. My own research found that Americans are in the throes of a "<a href="https://www.nationalreview.com/2021/07/american-men-suffer-a-friendship-recession/" target="_blank" rel="noopener">friendship recession</a>" with people reporting smaller social circles and fewer close friends. This rising tide of isolation is <a href="https://www.businessinsider.com/gen-z-most-accomplished-loneliest-generation-blame-parents-fault-2022-4">particularly acute among young people</a>: The time that Americans between the ages of 15 and 24 spend with friends has declined considerably over the past two decades, according to the surgeon general's report, from an average of 2.5 hours a day to just 40 minutes. It seems as if everything in modern life is conspiring to perpetuate the loneliness problem — from <a href="https://www.nbcnews.com/health/health-news/social-media-mental-health-anxiety-depression-teens-surgeon-general-rcna85575" target="_blank" rel="noopener">the design of our technology</a> to <a href="https://www.businessinsider.com/millennials-gen-z-blame-loneliness-on-housing-affordability-crisis-2023-4">where we build our homes</a>. We already know how addictive social media can be: Nearly one in three Americans reports being online "<a href="https://www.pewresearch.org/short-reads/2021/03/26/about-three-in-ten-u-s-adults-say-they-are-almost-constantly-online/ft_2021-03-26_alwaysonline_01/" target="_blank" rel="noopener">almost constantly</a>," according to the Pew Research Center, while <a href="https://penntoday.upenn.edu/news/social-media-use-increases-depression-and-loneliness" target="_blank" rel="noopener">a 2018 study by researchers</a> at the University of Pennsylvania showed that social media helps fuel feelings of loneliness. The latest development that threatens to make this loneliness crisis even worse is the rise of artificial intelligence. The release of OpenAI's ChatGPT in late 2022 has led to an explosion of interest in the potential to integrate AI-driven chatbots into our lives. <a href="https://www.theatlantic.com/ideas/archive/2023/04/ai-technology-productivity-time-wasting/673880/" target="_blank" rel="noopener">Derek Thompson</a>, a writer at The Atlantic, suggested that AI in its current manifestation is mostly a diversion, a waste of time. That may be true, but as previous technologies have shown us, it's crucial to take stock of the ways in which AI could shift our lives before it becomes ubiquitous.We've already seen how dependence on technology can weigh on our mental health, and now chatbots and other AI programs could further replace the critical social interactions that help us build community. Many Amercians already harbor this worry: A <a href="https://www.pewresearch.org/internet/2022/03/17/how-americans-think-about-artificial-intelligence/" target="_blank" rel="noopener">recent survey by the Pew Research Center</a> found that one of Americans' top concerns with AI is the technology's inherent "lack of human connection." Our time on earth is limited. While the convenience of AI could provide many benefits, it can't replace time spent with real, living people. <h2>What do we lose when we lose the little interactions?</h2>Humans have used technology to handle dangerous, monotonous, and menial tasks for generations. And there are definitely areas where AI chatbots could be useful — especially when it comes to business. Software developers are using AI to streamline the coding process, lawyers are using it to help draft legal briefs, and some businesses are experimenting with chatbots in customer-service roles. While the use cases could boost a business' bottom line, it's important to be clear-eyed about the problems that arise from using AI to replace routine social exchanges. No human interaction is meaningless. Small acts of kindness (or cruelty) can have far-reaching effects, and even relationships with people we don't know well can be valuable. Sociologists have dubbed these brief encounters as "<a href="https://www.journals.uchicago.edu/doi/abs/10.1086/225469" target="_blank" rel="noopener">weak social ties</a>" — a brief chat with a neighbor, a compliment from a stranger, or the barista at the local coffee shop asking how you've been. While the interactions often seem trivial, the benefits of these encounters are significant. Interviewing New York City residents living alone during the pandemic, the sociologists Eric Klinenberg and Jenny Leigh found that people missed being around "<a href="https://www.theatlantic.com/podcasts/archive/2023/05/social-infrastructure-public-space-community-relationships/674157/" target="_blank" rel="noopener">familiar strangers,</a>" an experience that gave them a sense of place and belonging. <a href="https://journals.sagepub.com/doi/abs/10.1177/0146167214529799?journalCode=pspc" target="_blank" rel="noopener">Studies have shown</a> that interacting with a wide array of people on a regular basis makes us happier, and my own research has shown that when it comes to friendships, <a href="https://www.americansurveycenter.org/yes-having-more-friends-is-better/" target="_blank" rel="noopener">more is better</a>. One of the profoundly important parts of developing social ties — whether strong or weak — is that they connect us to places and people we might not otherwise have access to. It might be a potential job opportunity or an introduction to a new community, such as a book club, religious congregation, or sports league. In a <a href="https://www.americansurveycenter.org/research/the-state-of-american-friendship-change-challenges-and-loss/" target="_blank" rel="noopener">2021 survey</a>, we found that nearly half of young adults made a close friend through their existing friendship network. Our current friendships beget new friendships.Using AI to automate these interactions, both trivial and more substantial, would deprive people of their mental and social benefits. The pandemic already gave us a glimpse at what happens when these ties fray: a rising number of <a href="https://www.theatlantic.com/politics/archive/2022/03/antisocial-behavior-crime-violence-increase-pandemic/627076/" target="_blank" rel="noopener">airport freakouts</a>, <a href="https://www.washingtonpost.com/education/2021/10/26/schools-violence-teachers-guns-fights/" target="_blank" rel="noopener">more frequent fighting in schools</a>, and a <a href="https://www.nytimes.com/2022/01/01/business/customer-service-pandemic-rage.html" target="_blank" rel="noopener">general increase</a> in antisocial behavior, among others. When we spend less time with each other, we lose practice in getting along in shared spaces. This is why AI is such a poor substitute for real-world interactions. We need to spend more time with each other.<h2>Artificial Intimacy</h2>The shift toward AI relationships isn't just a theoretical possibility: Some entrepreneurs and companies are already working to create chatbot-driven connections. Caryn Marjorie, a 23-year-old influencer with more than 1.8 million Snapchat followers, <a href="https://www.insider.com/influencer-creates-ai-virtual-girlfriend-herself-charges-dollar-minute-fortune-2023-5">recently released CarynAI</a> — an "immersive AI experience" featuring videos of Marjorie that she says provide a "virtual girlfriend" for those willing to pony up $1 per minute. According to Marjorie's site, the GPT-4-powered chatbot replicates Marjorie's voice and personality to the point that it feels like "you're talking directly to Caryn herself." While Marjorie stands to make a tidy profit from CarynAI — based on a recent beta test she estimated that the chatbot could generate $5 million in monthly revenue — she also says the goal in developing the AI avatar was to "cure loneliness" for her overwhelmingly male fan base. While it has garnered a lot of attention, CarynAI isn't the first attempt at providing people with companionship through an AI chatbot. Replika, another AI-chatbot friend marketed to people who are "lonely, depressed, or have few social connections" launched in 2017. The company behind Replika has over <a href="https://www.vice.com/en/article/93bqbp/can-you-be-in-relationship-with-replika" target="_blank" rel="noopener">10 million registered users</a>, and the chatbot receives millions of messages each week. Snap recently deployed its own AI chatbot, called My AI, aimed at supplementing social interactions on the app. In an interview with <a href="https://www.theverge.com/2023/2/27/23614959/snapchat-my-ai-chatbot-chatgpt-openai-plus-subscription" target="_blank" rel="noopener">The Verge</a>, Snap CEO Evan Spiegel said: "The big idea is that in addition to talking to our friends and family every day, we're going to talk to AI every day."One problem that's already apparent is the amount of time users are spending with chatbots like CarynAI. Although many are designed to wind down conversations after a designated period of time, there are no enforced limits, and Marjorie's manager<a href="https://www.washingtonpost.com/technology/2023/05/13/caryn-ai-technology-gpt-4/" target="_blank" rel="noopener"> told The Washington Post</a> that many fans are already spending hours each day sharing intimate thoughts and feelings with the bot.At least in the early iterations, AI chatbots appear to offer the next step in the curation of individualized experiences — another way to mold our world to our own preferences and mirror our own thoughts. Replika promises that it is "always here to listen and talk. Always on your side." But that's not a realistic model for most relationships, which should be built on a foundation of give-and-take and mutual obligation, rather than one-way devotion. In a review of Replika, Joshua Bote of the<a href="https://www.sfgate.com/tech/article/replika-san-francisco-ai-chatbot-17915543.php" target="_blank" rel="noopener"> San Francisco Chronicle</a> put it succinctly: "Replika, in its total fealty to its users, mostly serves as a vessel for users' wants and needs." The chatbot is unlikely to ask a user for a favor, much less encourage them to address destructive or selfish behavior.As I argued in a <a href="https://storylines.substack.com/p/will-artificial-intelligence-make" target="_blank" rel="noopener">recent newsletter</a>, this kind of one-way relationship is a poor substitute for traditional friendship:"The relationships that matter are formative. They change us. They provide us opportunities to practice forgiveness, patience, and kindness. The most valuable relationships are those that motivate us to become better. Relationships that do not require empathy and understanding rob us of the very things that make them so important."Most of us do not become better people through sheer force of will. Rather, our relationships give us reason to make difficult choices and changes. We make sacrifices for the people we love and who love us. In doing so, we can become the best version of ourselves. <h2>A problem for Gen Z</h2>Growing up, my brothers and I would often complain to my parents that we were bored. After explaining to us that this was not their problem, they would typically shoo us outside and we would wander the neighborhood. Sometimes we would bump into friends or talk to neighbors. If we were feeling ambitious, we would bike over to a few local shops. That world of casual interaction that my brothers and I grew up in is largely gone. Teens especially are spending far less time with each other out in the real world. In the late 1970s, <a href="https://theconversation.com/teens-have-less-face-time-with-their-friends-and-are-lonelier-than-ever-113240" target="_blank" rel="noopener">more than half</a> of high-school seniors saw their friends every day, but that dropped to only 28% in 2017. Today, young people can wander around the limitless expanse of the internet while parked in their basement or bedroom.Gen Z may be uniquely predisposed to seek out relationships with AI-generated avatars. Not only are they more comfortable using technology in this way, compared to previous generations, young adults are also participating less often in traditional social activities like having regular family dinners, attending religious services, or playing sports. In a 2021 study, we found that <a href="https://www.americansurveycenter.org/featured_data/few-gen-zers-grew-up-having-family-dinners/" target="_blank" rel="noopener">only 38% of Gen Zers</a> grew up having daily meals with their family. In contrast, more than three-quarters (76%) of baby boomers report that they had daily meals with their family when they were young. More structured social outlets, such <a href="https://storylines.substack.com/p/decline-of-youth-sports-drives-teen-loneliness" target="_blank" rel="noopener">as youth sports, are experiencing a rapid decline</a>.Young people are not only entering adulthood with fewer close friends than in the past, but fewer opportunities to forge new social connections. Other than social media, young adults have diminishing options to connect with their peers. But social media presents its own risks. The surgeon general recently said that social media represents a "<a href="https://www.hhs.gov/sites/default/files/sg-youth-mental-health-social-media-advisory.pdf" target="_blank" rel="noopener">profound risk of harm</a>" to adolescent mental health, which has worsened in recent years. <h2>A way AI could potentially make things better</h2>This doesn't mean there aren't opportunities for AI to enrich our lives in some ways, especially for people who face challenges that limit more traditional social opportunities. Brown University and the Hasbro toy company were <a href="https://www.brown.edu/news/2017-11-01/aries" target="_blank" rel="noopener">awarded a $1 million National Science Foundation grant</a> to develop AI pets to help seniors with everyday tasks, such as remembering to take medication. My colleague Brent Orrell recently <a href="https://www.aei.org/workforce-development/will-chat-tech-help-the-neuro-divergent-find-their-place-in-society/" target="_blank" rel="noopener">wrote</a> about how AI chatbots might allow neurodivergent people to better navigate challenging social dynamics. In researching new AI chatbots, I was struck by the missed opportunity for disconnected people to forge new connections. What if instead of substituting for human connections that people are lacking, AI chatbots could connect lonely people to each other in an attempt to foster personal connections? Imagine an AI chatbot that identifies people with similar interests or needs and then brokers an introduction and encourages real-world social interactions. Think of it as a mutual friend. Instead of using it to strip away users' humanity, AI technology should enhance social opportunities.While there may be upsides to the technology, my own studies on human connection and our social lives make me concerned about leaning on AI to avoid the unpleasant realities of human existence. No one goes through life without ever feeling lonely. It's a fundamental and universal human experience. The goal should not be to avoid these feelings, but to use these experiences to inform our decisions about how we want to live and what really matters to us.Artificial intelligence is poised to transform American society, but making things easier is not always an improvement, especially when it comes to social interactions. As Klinenberg, the sociologist, notes: "Efficiency is the enemy of social life." The solution to loneliness is not to develop increasingly clever social distractions, but to go outside and wander around. What we need now more than ever are safe and accessible public spaces for people to wander.<a href="https://www.businessinsider.com/author/daniel-cox" rel="">Daniel Cox</a> is director of the Survey Center on American Life and research fellow in polling and public opinion at the American Enterprise Institute.

Generation GPT: How AI technology will screw over young people

<ul class="summary-list"><li>ChatGPT has already passed the US Medical Licensing Exam, the test all doctors must take.</li><li>Now radiologists say it can also pass their specialist board exam.</li></ul>ChatGPT is becoming a great medical test taker.Its latest and most advanced version, ChatGPT-4, can already pass the <a href="https://www.insider.com/chatgpt-passes-medical-exam-diagnoses-rare-condition-2023-4#:~:text=US%20medical%20exam%20licensing%20questions%20correctly%20more%20than%2090%25%20of%20the%20time.">US Medical Licensing Exam</a> with flying colors. And now, it's moved one step closer to becoming a specialized physician. Well, sort of. On Tuesday, <a href="https://www.eurekalert.org/news-releases/988870?" target="_blank" rel="noopener">scientists announced</a> that the newest chatbot software from OpenAI can handily pass a Canadian or US-style radiology board exam – scoring more than ten points above the 70% passing threshold. But there's a catch: Because ChatGPT is only designed to process language (so far), the AI's radiology exam didn't include any images. Feels like kind of a major oversight for a branch of medicine concerned with diagnosing based on X-rays, MRIs, and other body pictures.<h2>ChatGPT did well on 'challenging' questions, but got some of the basics of radiology wrong</h2><div type="image" title="ct scan" height="1335" width="2000" msid="100286779" hostid="261" sizebytes="144744"></div>ChatGPT has shown itself to be <a href="https://www.businessinsider.com/list-here-are-the-exams-chatgpt-has-passed-so-far-2023-1#sommelier-examinations-8">a formidable test taker</a> — it's passed exams including the SAT, the bar exam, and even the challenging master sommelier tests. On the radiology exam, ChatGPT-4 delivered a passing score of 81%. The bot scored particularly well on higher-order thinking questions that require skills beyond just memory recall, like analysis, synthesis, and evaluation; ChatGPT-4 did well at describing imaging findings (85%), and applying concepts (90%). But it didn't do so great on some of the more straightforward questions on the test, and actually got 12 of those questions wrong that its predecessor, GPT 3.5, got right."We were initially surprised by ChatGPT's accurate and confident answers to some challenging radiology questions," study author Dr. Rajesh Bhayana, an abdominal radiologist at Toronto General Hospital, said in a <a href="https://www.eurekalert.org/news-releases/988870?" target="_blank" rel="noopener">press release</a>. "But then equally surprised by some very illogical and inaccurate assertions."<h2>Why ChatGPT is increasingly good at acing hard tests </h2>At its core, ChatGPT is simply designed to predict the best next word to spit out – in other words, it's <a href="https://www.businessinsider.com/everything-you-need-to-know-about-chat-gpt-2023-1#:~:text=Chat%20bots%20like%20GPT%20are%20powered%20by%20large%20amounts%20of%20data%20and%20computing%20techniques%20to%20make%20predictions%20to%20string%20words%20together%20in%20a%20meaningful%20way.">built to chat</a> you up. That chattiness can easily be harnessed to pontificate on exams, producing a very confident (but sometimes wrong) test taker. Brown computer science professor Ellie Pavlick, a natural language processing expert, says this issue isn't unique to chatbots. She's always struggled to tell whether students really grasp the concepts she's taught based on their written exam answers. "If you give some well-constructed language, it seems like you know, but maybe you do, maybe you don't," Pavlick said during a recent <a href="https://youtu.be/zVQVjljd3QA?t=3024" target="_blank" rel="noopener">ChatGPT roundtable at Brown University</a>. "This is a nice indication of why we kind of want to attribute much more knowledge and awareness to ChatGPT than it actually really has, because there's just something about well-constructed language that can really mask [poor] understanding."Doctors say that ChatGPT shouldn't be used to diagnose or treat patients, and should always have its medical accuracy checked by a person. But medical experts are also discovering how ChatGPT can be a nice tool for <a href="https://www.businessinsider.com/chatgpt-more-empathetic-than-doctors-study-save-time-bedside-manner-2023-5">improving a doctor's communication with their patients</a>. It is precisely because ChatGPT excels at banter that it is often regarded as more compassionate than hurried doctors.

After passing another medical exam, ChatGPT is moving one step closer to ChatMD

<ul class="summary-list"><li>This post originally appeared in the Insider Today newsletter.</li></ul>Hello! Welcome back to our Sunday edition, a roundup of some of our top stories. Ever wonder how much money you need to make to be considered middle <a target="_blank" href="https://www.businessinsider.com/upper-class-income-cutoff-wealth-middle-class-washington-dc-henrys-2024-5">class</a>? What about the <a target="_blank" href="https://www.businessinsider.com/upper-class-income-cutoff-wealth-middle-class-washington-dc-henrys-2024-5">incomes of the upper class</a>? We <a target="_blank" href="https://www.businessinsider.com/upper-middle-lower-class-income-wealth-poverty-new-york-census-2024-5">crunched the numbers</a>.On the agenda:<ul><li>How Oracle's biggest-ever acquisition <a target="_blank" href="https://www.businessinsider.com/oracle-cerner-health-larry-ellison-28-billion-deadly-gamble-veterans-2024-5">turned deadly</a>.</li><li>Robots and AI might just be <a target="_blank" href="https://www.businessinsider.com/conceivable-automating-ivf-fertility-robots-ai-startup-2024-5">the future of IVF</a>.</li><li>Millennials are <a target="_blank" href="https://www.businessinsider.com/millennials-retiring-early-fire-movement-financial-independence-millionaire-savings-work-2024-5">redefining the FIRE movement</a>.</li><li>Neighborhoods have really <a target="_blank" href="https://www.businessinsider.com/america-dumb-neighborhood-names-real-estate-rebrands-cities-apartments-development-2024-5">dumb names these days</a>.</li></ul>But first: OpenAI's in the spotlight (again).If this was forwarded to you, <a target="_blank" href="https://www.businessinsider.com/subscription/newsletter/insider-today-sunday-edition">sign up here</a>. Download Insider's app <a target="_blank" href="https://www.insider.com/app">here.</a><h2 id="e315fe86-3322-4db2-90d7-81d7d862c4a8">This week's dispatch</h2>When OpenAI's <a target="_blank" href="https://www.businessinsider.com/openai-gpt4o-launch-ai-new-best-friend-sam-altman-chatgpt-2024-5">GPT-4o </a>started chatting during the AI startup's <a target="_blank" href="https://www.businessinsider.com/openai-event-live-updates-sam-altman-announcement-chatgpt-news-2024-5">livestream event</a>, I was immediately reminded of "Her," the Spike Jonze movie from a decade ago. I wasn't the only one. Scarlett Johansson, who voiced the AI Samantha in that movie, said last week that <a target="_blank" href="https://www.businessinsider.com/scarlett-johansson-angry-chatgpt-voice-declined-offer-openai-sam-altman-2024-5">she had hired lawyers</a> after she heard GPT-4o using a voice that sounded "eerily similar" to her own. This was after Sam Altman tried to hire her to <a target="_blank" href="https://www.businessinsider.com/scarlett-johansson-angry-chatgpt-voice-declined-offer-openai-sam-altman-2024-5">voice the AI bot, and she declined. </a>OpenAI paused the voice, saying a different actress was behind it, and that the voice was cast before Altman reached out to Johansson. That one of the AI voices sounds a great deal like a Hollywood star who declined to work with OpenAI is <a target="_blank" href="https://www.businessinsider.com/openai-scarlett-johansson-voice-defense-sam-altman-washington-post-chatgpt-2024-5">purely coincidental</a>. OK!It's not a good look for OpenAI, which is facing continued questions over its truthfulness right as it tries to become a <a target="_blank" href="https://www.businessinsider.com/openai-scarlett-johansson-sam-altman-chatgpt-disaster-why-2024-5">fixture in all of our lives</a>.Alphabet CEO Sundar Pichai said OpenAI might have <a target="_blank" rel="noopener" href="https://www.businessinsider.com/sundar-pichai-openai-breached-youtube-terms-train-sora-2024-5">breached YouTube's terms</a> when training its Sora model. Authors are <a target="_blank" rel="noopener" href="https://www.businessinsider.com/openai-destroyed-ai-training-datasets-lawsuit-authors-books-copyright-2024-5">taking legal action</a> over breach of copyright. There have been multiple <a target="_blank" rel="noopener" href="https://www.businessinsider.com/sam-altman-openai-glory-days-ending-problems-chatgpt-2024-5">high-profile departures</a> in recent months.Six months on from his brief ouster as CEO, Altman is under pressure again.<div type="image" title="Larry Ellison" height="1000" width="2000" msid="110440678" hostid="261" sizebytes="197650"></div><h2 id="13dc2e63-40ad-4acd-8b2a-94db0cc7a6b6">Oracle's deadly gamble</h2>Larry Ellison bet $28 billion that he could revolutionize healthcare. With the purchase of medical-records company Cerner, the billionaire planned to pump its vast trove of medical data into Oracle's AI models and develop an electronic health record system of the future. So far, it's been a disaster.<a target="_blank" href="https://www.businessinsider.com/oracle-cerner-health-larry-ellison-28-billion-deadly-gamble-veterans-2024-5">Inside Ellison's disastrous bet on healthcare</a>.<div type="image" title="Conceivable Life Sciences" height="560" width="1118" msid="110440680" hostid="261" sizebytes="45432"></div><h2 id="320be494-865e-4ab4-b8d6-ef805f065a02">The future of fertilization</h2>Conceivable Life Sciences, a startup in Guadalajara, Mexico, is building robots and AI models to automate crucial parts of IVF. Though they're just in the prototype stage, the robots can suck sperm into needles and place them into eggs.The cofounders say automation will challenge an industry rife with arbitrary decision-making — and it could wind up saving patients tens of thousands of dollars. <a target="_blank" href="https://www.businessinsider.com/conceivable-automating-ivf-fertility-robots-ai-startup-2024-5">Here's how it works</a>.<div type="image" title="Illustration of a man distracted at work on a gold course." height="2000" width="4000" msid="110440682" hostid="261" sizebytes="199566"></div><h2 id="c00b4c10-ed66-4e49-aeee-9762feccf0de">FI, not RE</h2>For some millennials, the FIRE movement — financial independence, retire early — isn't just about riding off into the sunset early. A portion of the generation is focused more on financial independence than cutting out work completely. The result means they can create their own version of life after work.<a target="_blank" href="https://www.businessinsider.com/millennials-retiring-early-fire-movement-financial-independence-millionaire-savings-work-2024-5">Millennials throw some cold water on FIRE</a>.Also read:<ul><li><a target="_blank" href="https://www.businessinsider.com/early-retiree-shares-money-moves-strategies-achieve-financial-independence-fire-2024-5">An early retiree shares 3 money moves he made to save seven figures, quit work at 48, and still live comfortably: 'You don't have to live on ramen'</a></li><li><a target="_blank" href="https://www.businessinsider.com/early-retiree-saved-too-much-in-401k-financial-independence-strategy-2024-5">A 50-year-old retiree says his 'biggest mistake' was saving too much in his 401(k)</a></li><li><a target="_blank" href="https://www.businessinsider.com/how-an-early-retiree-invested-his-money-to-create-wealth-2024-5">An early retiree shares the advice he received in his early 20s that helped him retire at age 48</a></li></ul><div type="image" title="Street sign that says &quot;Welcome to WeLoNoBroSoWa&quot;" height="650" width="1300" msid="110440684" hostid="261" sizebytes="54184"></div><h2 id="14d71008-b99c-492f-91ef-3354a5de5b3a">Welcome to DumBro</h2>Cities on the rise have a weird trend these days: Really dumb neighborhood names. Take Denver, where you can find RiNo, SoBo, LoDo, and LoHi. Two-syllable names for neighborhoods used to be an oddity.But now it's become a full-blown epidemic as builders try to redevelop portions of cities into their image. And it's not just an innocent name game. Oftentimes it's about erasing the history — and pushing residents out — of the original neighborhood.<a target="_blank" href="https://www.businessinsider.com/america-dumb-neighborhood-names-real-estate-rebrands-cities-apartments-development-2024-5">The great neighborhood renaming</a>.This week's quote:"I channeled a thick-skinned character who thrived in a man's world."— Lindsay MacMillan, a former Goldman Sachs VP who was <a target="_blank" href="https://www.businessinsider.com/women-leaving-goldman-sachs-boys-club-why-i-quit-2024-5">driven out by the company's "boys only" culture</a>.<h2 id="a4a20687-c2b5-4eca-b5f4-c89728a89254">More of this week's top reads:</h2><ul><li>These are the top 15 <a target="_blank" href="https://www.businessinsider.com/united-states-best-places-to-live-cheaper-homes-quality-life-2024-5">places to live in the United States</a>.</li><li>Baby boomers have been America's <a target="_blank" href="https://www.businessinsider.com/us-recession-baby-boomer-consumer-spending-downturn-retirement-hard-landing-2024-5">secret weapon for fending off recession</a>.</li><li>Homes on Nantucket are falling into the ocean — but the <a target="_blank" href="https://www.businessinsider.com/nantucket-homes-wealthy-falling-into-ocean-erosion-climate-2024-5">wealthy won't stop buying them</a>.</li><li>Companies are "unbossing" the workplace. <a target="_blank" href="https://www.businessinsider.com/companies-are-unbossing-workplaces-and-millennials-are-at-risk-2024-5">Millennial managers are getting axed</a>.</li><li>How Khaby Lame, the world's most-followed TikToker, <a target="_blank" href="https://www.businessinsider.com/khaby-lame-business-biggest-tiktok-creator-in-the-world-2024-5">built his business</a>.</li><li>Deleting Slack and Gmail helped one woman <a target="_blank" href="https://www.businessinsider.com/deleting-gmail-slack-phone-entrepreneur-parent-2024-5">become a better parent — and worker</a>. </li><li>Google AI said to put glue in pizza — so <a target="_blank" rel="noopener" href="https://www.businessinsider.com/google-ai-glue-pizza-i-tried-it-2024-5">I made a pizza with glue and ate it</a>.The Insider Today team: <a target="_blank" href="https://www.businessinsider.com/author/matt-turner">Matt Turner</a>, deputy editor-in-chief, in New York. <a target="_blank" href="https://www.businessinsider.com/author/jordan-parker-erb">Jordan Parker Erb</a>, editor, in New York. <a target="_blank" href="https://www.businessinsider.com/author/dan-defrancesco">Dan DeFrancesco</a>, deputy editor and anchor, in New York. <a target="_blank" href="https://www.businessinsider.com/author/lisa-ryan">Lisa Ryan</a>, executive editor, in New York.</li></ul>

Neighborhoods have really dumb names these days

<ul class="summary-list"><li>John Pasmore founded Latimer, a startup that's weaving Black data into AI model outputs.</li><li>He's building partnerships with HBCUs and Black media producers.</li></ul>On a recent cold, rainy afternoon in downtown Brooklyn, I met John Pasmore, the CEO of <a target="_blank" href="https://www.businessinsider.com/latimer-ai-api-launch-bias-detection-tools-llm-2024-2">AI startup Latimer</a>.He's a Black founder who is trying to accomplish something that seems almost impossible these days: <a target="_blank" href="https://www.businessinsider.com/ai-art-generators-dalle-stable-diffusion-racial-gender-bias-ceo-2023-3">Making AI less biased</a>. Just don't call it "BlackGPT." He's not a fan of the moniker, which has somehow attached itself to his project.After years of preaching about AI ethics, the tech industry has piled into a race to build the most powerful AI models. Making sure these products reflect the diversity of different cultures has taken a back seat to <a target="_blank" href="https://www.businessinsider.com/ai-data-centers-energy-cloud-microsoft-chatpgt-amazon-google-2023-7">energy-sucking training runs</a>, obsession with benchmarks, and <a target="_blank" href="https://www.businessinsider.com/the-scope-of-sam-altmans-future-ai-empire-open-ai-2024-3">endless pitches about AGI </a>and the coming <a target="_blank" href="https://www.businessinsider.com/ai-work-chatgpt-workers-automation-wharton-professor-2023-6">productivity revolution</a>.If all this seems overwhelming, Pasmore doesn't show it. For our interview, he chatted freely while tucked away on a couch in a bar at a Brooklyn coworking space, sipping Maker's Mark neat.In fact, the glaring shortcomings of AI models, and a relative lack of other founders working on these problems, attracted Pasmore.AI has well-documented blindspots when it comes to Black culture. Outputs are biased because <a target="_blank" href="https://www.businessinsider.com/why-facial-recognition-technology-might-not-recognize-you-2024-3">models are trained on data that is mostly white, male, and Western</a>. Pasmore realized he might be able to make a better alternative, and founded Latimer in May 2023 to try."I just didn't think that people on Twitter asking one of the giant tech companies to fix Black history was going to work out," Pasmore told me. "I didn't think they would be responsive to that, and that it was something we could do ourselves."<h2>Navigating two worlds</h2>Pasmore is well positioned for this moment, with a background that gives him unique perspectives on AI's promise and its failings.He grew up in a majority-white, middle-class neighborhood on Long Island. His father came to the US from Jamaica and owned a gas station. His mother, from Trinidad, worked as a nurse for PanAm Airlines when it offered in-house medical care to airline staff."Based on my background and childhood I can kind of exist in and navigate both worlds fairly easily," Pasmore said.In his 20s, he moved to Texas and worked at Chemical Bank, which ended up as part of JPMorgan Chase. Rather than climb the rungs of corporate finance, Pasmore decided to go in another direction."I jumped off the corporate path fairly early on," he recalled. "My first business, Intrigue Magazine, was kind of a youth-culture, city-guide publication."<h2>From Intrigue to AI</h2><div type="image" title="Russell Simmons" height="2053" width="3000" msid="108858209" hostid="261"></div>Intrigue was eventually sold for an undisclosed amount. By 1995, Pasmore and hip-hop mogul <a target="_blank" rel="noopener noreferrer" class="c-link" href="https://www.businessinsider.com/russell-simmons-officially-launches-youtube-network-all-def-digital-2013-11">Russell Simmons</a> were building media brand OneWorld. From there, Pasmore joined Paris-based Trace TV as an EVP. But it wasn't long before the founder bug came biting, and Pasmore launched Voyage.tv in 2007. That venture-backed video travel platform was sold to Next 1 Interactive in 2013 for $10 million.Then came advisory roles at companies like Strides AI, where he worked to diversify corporate boards. He also put in a seven-year stint as managing partner at messaging platform Yestime.He always felt he'd missed out on being a technical founder and became increasingly interested in AI. In 2019, he went back to school and earned a BA in computer science from Columbia University with further training in machine learning from Cornell.Along the way, he also returned to his finance roots and joined private equity firm TRS Capital as a partner in 2021.<h2>Esther backs Latimer</h2><div type="image" title="Esther Dyson" height="1066" width="1599" msid="108858212" hostid="261"></div>Pasmore saw that the conversation around AI, and the need for inclusive tools, had reached a new inflection point.By this stage, Pasmore had intimate experience with a common conundrum among Black tech entrepreneurs: How to catch the attention of venture capitalists and raise money for your idea."The expectation on the VC front is that you're able to go out to your friends and family and raise a bunch of money," Pasmore told BI. "That is a hurdle if your set of friends and family is not in a position to do that."So, Pasmore tapped his network. He'd known prominent angel investor <a target="_blank" href="https://www.businessinsider.com/investor-esther-dyson-is-so-into-health-tech-she-actually-wears-monitoring-gadgets-2011-12">Esther Dyson</a> for more than a decade and pitched her on his vision for Latimer.She quickly became Latimer's first investor because she saw the startup's potential to address AI's troubling tendency to ignore Black history and culture."AI won't solve all of these problems, but it will do what we as humans ask it to do. Use Latimer, ask it to find the history that was hidden," Dyson told BI. "An AI hallucination is a hypothesis and the only way to prove one or disprove one is to go search out the facts."Dyson, daughter of <a target="_blank" href="https://www.businessinsider.com/legendary-physicist-freeman-dyson-death-age-2020-2">famed scientist Freeman Dyson</a>, has backed successful startups, including Flickr and Square, while chairing the influential Electronic Frontier Foundation."The value of AI isn't the nodes, it's in the edges and the relationship between the kinds of data," she said. "The big business model is going to be quality assurance for the information supply chain."<h2>A different approach to AI training data</h2>Legitimizing information sources that form the basis of training data for AI models will become increasingly important, and it's something Pasmore has spent a lot of time thinking about.Rather than <a target="_blank" href="https://www.businessinsider.com/ai-killing-web-grand-bargain-2023-8">web crawlers sucking up information from the internet</a>, which makes it hard to give <a target="_blank" href="https://www.businessinsider.com/joy-buolamwini-unmasking-ai-civil-biometric-creative-rights-dei-issue-2024-2">compensation, consent, control, and credit</a>, Latimer's data strategy is different.It centers on partnerships with historically Black colleges and universities, and licensing deals with Black media companies and other similar content producers.The lead content czar for Latimer's database is Temple University's Molefi Kete Asante, who created the school's first African American studies Ph.D. program."Black and Brown publishers want to both financially benefit from having produced this content, but also don't want the content to be so siloed that you can only get accurate information from one place," Pasmore said.Latimer has also partnered with Morgan State University to develop a method for detecting bias in AI-generated text. This method is baked into a suite of business offerings Latimer is set to launch soon. In February, the company announced it was <a target="_blank" href="https://www.businessinsider.com/latimer-ai-api-launch-bias-detection-tools-llm-2024-2">launching an API that can detect bias in text</a>.<h2>How Latimer's tech works</h2>Now, just under a year after Latimer started, Pasmore and his team of eight employees are busy building out its specialized datasets and integrating this with OpenAI GPT models and alternatives such as open-source model provider <a target="_blank" href="https://www.businessinsider.com/mistral-in-talks-to-raise-funding-at-2-billion-valuation-2023-11">Mistral AI</a>.Latimer works by tapping OpenAI's GPT-3.5 for its baseline capabilities. Before Latimer answers queries, it pulls in information from a trove of data on Black culture and history through a process known as retrieval-augmented generation, or RAG.Latimer is among a growing list of companies building new tools atop foundational AI models using RAG, including search-engine startup <a target="_blank" href="https://www.businessinsider.com/perplexity-raising-new-funding-generative-ai-2024">Perplexity AI</a>.Pasmore's startup has been differentiated from rivals by being labeled the <a target="_blank" href="https://www.businessinsider.com/latimer-ai-api-launch-bias-detection-tools-llm-2024-2">"BlackGPT."</a>While catchy, Pasmore is not a fan of the nickname. That's partly because it's too close to ChatBlackGPT, a product developed by Erin Reddick to highlight Black culture and history.Of course, if AI models actually reflected diversity accurately, there'd be no need for such a name. Pasmore wants to make this a reality someday.Correction: March 28, 2024 — An earlier version of this story misstated a relationship between Intrigue and OneWorld. These businesses are separate entities.

In an ocean of white data, this founder is trying to make AI less biased. Just don't call it 'BlackGPT.'

The hidden wave of employees using AI on the slyFor the most part, Blake doesn't mind his job as a customer-benefits advisor at an insurance company. But there's one task he's always found tedious: scrambling to find the right medical codes when customers call to file a claim. Blake is evaluated in part on the amount of time he spends on intake calls — the less, the better — and the code-searching typically takes him two or three minutes out of a 12-minute call.Then he discovered that Bing Chat, Microsoft's AI bot, could find the codes in mere seconds. At a call center, a productivity gain of 25% or more is huge — the kind that, if you told your boss about it, would win you major accolades, or maybe even a raise. Yet Blake has kept his discovery a secret. He hasn't told a soul about it, not even his coworkers. And he's kept right on using Bing to do his job even after his company issued a policy barring the staff from using AI. Bing is his secret weapon in a competitive environment — and he isn't about to give it up."My average handle time is one of the lowest in the company because I'm leveraging AI to accelerate my work behind their back," says Blake, who asked me not to use his real name. "I'm totally going to take advantage of it. This is part of a larger way of making my life more efficient." Since ChatGPT came out last November, employees in corporate America have responded in a variety of ways. Some have fought back against the use of <a href="https://www.businessinsider.com/chatgpt-ai-technology-end-of-coding-software-developers-jobs-2023-4">AI</a>, worried about their job security. Others are waiting for their companies to train them in how to use the new technology. And then there are employees like Blake — early adopters who are quietly using AI to do their jobs faster and better, even if it means violating company policy. Call it CheatGPT — a move that gives employees who are willing to bend or even break the rules a hidden advantage over their tech-averse coworkers.Those secretly using AI on the job — experts call it "shadow IT" — appear to be legion. Back in January, even before rival tools like Bing Chat and Google's Bard were released, two-thirds of ChatGPT users surveyed by the social network Fishbowl said they were deploying the technology <a href="https://www.fishbowlapp.com/insights/70-percent-of-workers-using-chatgpt-at-work-are-not-telling-their-boss/" target="_blank" rel="noopener">on the sly</a>. That shouldn't come as a surprise, given the power of AI to boost productivity. In one study, AI made computer programmers 56% <a href="https://arxiv.org/pdf/2302.06590.pdf" target="_blank" rel="noopener">faster at coding</a>. In another, employees <a href="https://economics.mit.edu/sites/default/files/inline-files/Noy_Zhang_1.pdf" target="_blank" rel="noopener">completing writing tasks</a> were 37% faster when they were assisted by AI. In many cases, those who use the new tool get an immediate leg up at work."From decades of research in innovation studying everyone from plumbers to librarians to surgeons, we know that, when given access to general purpose tools, people figure out ways to use them to make their jobs easier and better," Ethan Mollick, a professor of management at the Wharton School who studies AI, <a href="https://www.oneusefulthing.org/p/detecting-the-secret-cyborgs" target="_blank" rel="noopener">recently observed</a>. But with technology like ChatGPT, he noted, employees aren't telling their companies about their discoveries. Instead, they've become "secret cyborgs, machine-augmented humans who keep themselves hidden." The race by employees to use AI — even if it means doing so in secret — is the opposite of what usually happens when a new technology arrives in the workplace. When a company implements new software, HR and IT typically spend months nagging everyone to use it, and employees comply only begrudgingly. This time, employees are rushing to use AI before their employers are ready. Why the flip?It's very much in an employer's interest to have more productive workers. But given the risks that accompany AI, most companies have been unwilling to give workers a green light. Some, like Blake's insurance company, fear AI platforms might gain access to sensitive customer information, which businesses are legally obliged to protect. Others worry that employees will inadvertently <a href="https://mashable.com/article/samsung-chatgpt-leak-details" target="_blank" rel="noopener">divulge trade secrets</a> in their prompts, or rely on <a href="https://www.businessinsider.com/ai-chatbot-chatgpt-google-microsofty-lying-search-belief-2023-2">error-prone responses</a> from a chatbot without checking the machine's work. A recent survey conducted by the research firm Gartner found 14% of companies had issued a blanket ban on the use of chatbots."A lot of organizations are trying to figure out what to do," says Eser Rizaoglu, a senior director analyst at Gartner. "I liken it a little bit to COVID, when no one had a playbook and everyone was just trying to figure it out as they go." In Gartner's survey, 35% of companies said they hadn't finalized their AI guidance yet, and another 18% said they didn't plan on issuing any guidance at all.Confusion about AI is rampant. "Employers are calling us," says Alex Alonso, the chief knowledge officer at SHRM, a trade association for HR professionals. "We're probably getting somewhere between 30 to 50 calls a week about that. 'What do I do with AI in the workplace?' And specifically, 'How do I deal with people who want to use it even though we haven't built a policy around it?'"Employees, for their part, aren't waiting for employers to catch up. Many are happily deploying AI to get ahead on the job — and to knock off early. A software engineer I'll call Roberto, who works at one of the largest retailers in the US, discovered that ChatGPT could save him as many as 15 hours a week on certain coding tasks. But instead of using that extra time to do more work, he's been studying for a class he's taking on the side."I'm remote," he says. "They can't tell when I'm working and when I'm not. I can ChatGPT my way through solving something in a couple hours, and then just take the rest of the day off and they won't know."Roberto hasn't told his coworkers or his boss about his use of AI. And even if his company decides to ban ChatGPT, he intends to keep right on using it. "If you use it in secret, you have an advantage over the people who aren't using it," he says. "So why talk about it? Why bring it up? I don't want to rock the boat."Others have stumbled onto AI as a sort of secret mentor. Luke, a software engineer who's new to the career, has a coworker he often turns to when he has questions. Then, a few months ago, he got stuck on a problem at a moment when the coworker was unavailable. In a panic, he asked ChatGPT for help — et voilà. "It gave me this beautiful skeleton code," he says. "I was like: Wow, this makes so much sense. Now I just have to fill in the gaps." Luke doesn't know whether his employer is OK with him using ChatGPT, since it hasn't issued an official policy, and he's not about to ask. He's kept on consulting the bot on the side, sometimes saving himself an entire day of work. And because he's getting more done, he feels less anxious in his weekly status meetings with his boss, who doesn't know what's behind his newfound productivity.<div type="image" title="Person looking at a phone with AI apps like ChatGPT" height="3960" width="5281" msid="102338766" hostid="261"></div>A few weeks ago, Luke decided to come clean with his coworker. He was nervous about it. "I wasn't sure what she'd say," he recalls. "But when I told her, she was like, 'Oh, that's funny,' and admitted that she uses it too. I think the unspoken thing is that everyone is using it." All this secret AI use is made much easier by remote work. "I have zero fear of getting caught," says Blake, who uses his personal computer to access Bing Chat. "There's no way they'll find out. I get so much privacy working from home." Even in an office, all employees need to do is pull up ChatGPT or Bing on their phones — the same way they check Facebook or Twitter when their companies block access to social-media sites. Employers can ban AI all they like, but there's no way they can stop it.If you're someone who's not using AI on the job, all this might strike you as unfair. But that's the reality at many workplaces right now: The failure of companies to adapt to the sudden emergence of tools like ChatGPT is creating a kind of AI inequality. Those who use it gain a quantifiable advantage over those who shy away from it. They're able to do more and better work, setting themselves up for the raises and promotions that might have gone to others. Or they're getting away with working less while everyone else is forced to toil away. By failing to create clear guidance on AI, companies are effectively empowering the covert users at the expense of everyone else."Let's say GPT was good enough to replace you as a writer," Mollick, the Wharton professor, told me a few months back. "It's going to be a long time before Insider and its parent companies are able to figure out how to take advantage of that in a useful way. So now is the time for you as a programmer, you as a writer, to do 10 times the work secretly as everybody else — because your company hasn't caught up."It doesn't have to be this way. By embracing AI, companies can create a level playing field for all employees. They can also take the productivity benefits being discovered by the stealth GPT users and scale them across entire teams and departments. But to do that, bosses need their secret AI users to stop being so secretive. That means coming up with creative ways to incentivize and reward employees who find good use cases for chatbots. "Think cash prizes that cover a year's salary," Mollick suggests. "Promotions. Corner offices. The ability to work from home forever. With the potential productivity gains possible due to large language models, these are small prices to pay for truly breakthrough innovation."It makes no sense for companies to shun a technology that many employees are eager to use — and that makes them more efficient and productive. Before long, smart employers won't be banning GPT use — they'll be celebrating it. In a sense, employees like Blake and Luke and Roberto aren't just getting an advantage at their current jobs. They're getting a résumé-enhancing jump on their future jobs. Eventually, even the most AI-hesitant companies will figure out how to let their employees use tools like ChatGPT — and at that point, they'll actively be seeking to hire workers who possess AI-savviness."Once companies feel safe using it, they're going to start to want people who are trained in it," Roberto says. "You'll start seeing on LinkedIn: 'Job applicants must have experience with prompt engineering.'" Sooner or later, using chatbots on the job won't be AI and lie. It'll be AI or die.<a href="https://www.businessinsider.com/author/aki-ito">Aki Ito</a> is a senior correspondent at Insider.

'I take the rest of the day off': How employees are secretly using AI to duck out early

<ul class="summary-list"><li>Biotech and pharma companies are evaluating how AI could make their businesses better.</li><li>The CEO of Eli Lilly said he expects AI to massively change the productivity of the workplace.</li><li>The pharma giant is investing in AI, machine learning, and automation.</li></ul>The pharma-and-biotech industry has already begun to use <a href="https://www.businessinsider.com/artificial-intelligence">artificial intelligence</a> to improve how it operates, develops drugs, and ultimately <a href="https://www.businessinsider.com/ai-digital-workers-new-mexico-transform-workplace-chatgpt-medicaid-automate-2023-5">makes things easier for employees.</a>A handful of biotech companies are <a href="https://www.businessinsider.com/list-top-ai-biotech-companies-recursion-schrodinger-abcellera">testing AI-developed drugs in people.</a> Meanwhile, digital-health companies, providers, and insurers are <a href="https://www.businessinsider.com/inside-the-healthcare-industrys-sprint-to-unleash-chatgpt-2023-5">grappling with how to use technologies including ChatGPT</a> to speed up tasks such as assessing patients and completing medical notes, while still maintaining the safety and privacy of their patients. According to David Ricks, the CEO of the pharma giant Eli Lilly, the technology has the potential to <a href="https://www.businessinsider.com/deepmind-alphafold-ai-origin-story-impact-on-biotech-2023-2">upend the industry.</a> Eli Lilly is <a href="https://www.lilly.com/discovery/clinical-development-pipeline" target="_blank" rel="noopener">developing dozens of drugs through clinical trials</a> and expects to bring in more than <a href="https://investor.lilly.com/static-files/1df8cab7-5fa8-4eb4-b72d-a93efb1dac56" target="_blank" rel="noopener">$30 billion in revenue</a> this year.Ricks told Insider that AI is "one of the most exciting technological moves" he's seen in a long time."I can only think of two other things in my adult life that would compete with it," Ricks said. "One of them was an iPhone, and another was when we first started to visualize the internet."Eli Lilly has a <a href="https://markets.businessinsider.com/stocks/lly-stock">market value of more than $420 billion</a>, and sells blockbuster treatments for diabetes and cancer. It's already begun to invest in a host of projects focused on AI. A spokesperson for the company said that Lilly is investing in artificial intelligence and machine learning in areas including drug discovery, natural-language generation, robotic-process automation, and chatbots.The goal is to grow what Lilly calls its "digital worker-equivalent workforce," a concept that the company says helps quantify the hours saved by using technology instead of human labor. Lilly said that its efforts, which began in 2022 and now span more than 100 projects, are equivalent to around 1.4 million hours of human activity, or around 160 years of 24/7 work.Lilly told Insider its goal is to bring this number to 2.4 million hours, or about 274 years, by the end of the year. Lilly declined to say how much it's invested in the initiative. <h2>Three ways Lilly wants to use AI </h2>Ricks said he sees three main ways Lilly and the larger biopharma space could use AI.Initially, the technology could carry out the first mundane steps in tasks such as contract production or the rote parts of administrative work."I think the most near-term application of these isn't going to be to replace whole people's roles, but rather augment human productivity, and starting with the easiest problems," he said. "I think it'll happen quickly."Second, in an industry with intense regulatory oversight, AI could also help automate repetitive business processes."That's a great target for some of the textual generative AI tools like <a href="https://www.businessinsider.com/everything-you-need-to-know-about-chat-gpt-2023-1">Chat GPT</a>, because we can teach it what to do on our data, and it can help people produce much faster documents that aren't actually creating value for the patient — they're contributing to the regulatory system," Ricks said. The final use case would be <a href="https://www.businessinsider.com/list-top-ai-biotech-companies-recursion-schrodinger-abcellera">drug development</a>, according to Ricks. An AI model could produce ideas based on a data set that chemists likely wouldn't have been able to see or visualize. In May, Lilly <a href="https://www.prnewswire.com/news-releases/xtalpi-announces-collaboration-with-lilly-using-ai--robotics-to-uncover-first-in-class-therapeutics-301837142.html" target="_blank" rel="noopener">announced</a> a $250 million partnership with pharmaceutical-technology company XtalPi to uncover new potential drugs using AI."In a discovery process, you want to funnel wide," Ricks said. "In the past, perhaps humans would just think of the things they already knew about. The machine doesn't. It just knows about everything that was there and it comes up with constructs that humans just don't."That construct is far from becoming an actual drug, Ricks said, but what it does is give the chemist a starting point, a "new white space," that they can use to eventually develop a new drug. "It's a breakthrough and I think it'll change the productivity of the workplace massively, and people will spend time on more interesting and valuable things as a result in the near term," he said.

The CEO of pharma giant Eli Lilly shares 3 ways AI could transform his industry

<ul class="summary-list"><li>The rise of generative AI in China has led to people try to recreate their loved ones with the tech.</li><li>Using old photos, recordings, and messages, they're training chat programs to imitate the dead.</li></ul>In 2020, a young Chinese software engineer in Hangzhou chanced upon an essay about lip syncing technology. Its premise is relatively simple — using a computer program to match a lip movements with speech recordings.But his grandfather, who died nearly a decade earlier, came to mind."Can I see Grandpa again using this technology?" Yu Jialin asked himself.His journey to recreate his grandfather, <a href="https://www.sixthtone.com/news/1012642" target="_blank" rel="noopener">documented in April by investigative journalist Tang Yucheng for the state-owned magazine Sixth Tone</a>, is one of several accounts now surfacing in China of people using artificial intelligence to resurrect the dead.Mixing an assortment of emerging AI technologies, people in the country have been building chat programs — known as griefbots — with the personalities and memories of the deceased, hoping for a chance to speak to their loved ones again.For Yu, they presented a chance to speak his final words to the man who helped raise him.The software engineer, now 29, told Tang that he was 17 when his grandfather died.He still regrets two instances when he was harsh to his grandfather. Yu yelled at the older man for interrupting a gaming session once, and on another occasion told his grandfather to stop picking him up from school, Tang reported.His family stopped mentioning his grandfather after he died, he told Tang. "Everyone in the family was trying their best to forget Grandpa rather than remember him," Yu said.<h2>The Griefbot rides the ChatGPT craze</h2>The griefbot concept has <a href="https://www.businessinsider.com/eternime-and-replika-giving-life-to-the-dead-with-new-technology-2018-11">been trialed for years</a> — largely as AI-powered programs that learn how to mimic human beings through their memorabilia, photos, and recordings. But generative AI's rapid advancement in the last year has pushed the power and accessibility of griefbots to a whole new level.Older models required vast sets of data. Now, laymen or lone engineers like Yu can feed language models with tidbits of a person's past, and recreate almost exactly how they look, speak, and think. "In today's technology, you don't need too many samples for an AI to learn the style of a person," Haibing Lu, an information and analytics professor at Santa Clara University, told Insider.Systems like ChatGPT, the popular text-based program that closely imitates human speech, have already learned how most people naturally speak or write, said Lu, whose research focuses on AI.<div type="image" title="Generative AI models have progressed to the point where they can even create entire Manga novels, like &quot;Cyberpunk: Peach John,&quot; which hit shelves in Japan in March." height="2828" width="4242" msid="100379875" hostid="261" sizebytes="576442"></div>"You only need to tweak the systems a little bit in order to loosely get a 99% similarity to your person. The stark differences will be minimal," Lu said.For Yu to teach his AI model what his grandfather was like, he retrieved a trove of old letters from his grandmother. She'd exchanged them with Yu's grandfather when they were young, and they revealed a side to the man that even Yu hadn't glimpsed as a child, he told Tang.The software engineer dug up photos and videos shot more than a decade ago, and found text messages his grandfather sent him, Tang reported.Yet even given weeks of testing and training, the tech has a long way to go if humans expect something akin to <a href="https://www.netflix.com/title/70264888" target="_blank" rel="noopener">Black Mirror's robot replicas.</a> Yu's bot was clearly limited, and took 10 minutes to respond to each prompt, Tang reported."Hey, Grandpa. Guess who I am?" Yu asked the program at one point.Grandpa delivered a generic response."Who you are is not important at all. Life is a beautiful miracle," the bot wrote back, according to Tang.But as Yu fed the AI with more information about his grandfather, it started to show a more accurate representation of the man's habits and preferences. For example, it remembered his grandfather's favorite show, he told Tang."Happy Teahouse went off the air," Yu told the chatbot."That's a shame. The show I want to watch the most is no longer available. I would have liked to watch a few more episodes," the grandfather bot replied.That was the moment when Yu felt he'd gotten somewhere, he told Tang. The program was eventually sophisticated enough that Yu felt confident he could show his work to his grandmother. She watched silently as her late husband responded to her questions, then thanked her grandson, stood up, and left the room.Yu told Tang his grandmother needed the chatbot to process her emotions and mourn. "Otherwise, why would she thank me?" he said.As for himself, he declined to share his intimate conversations with his grandfather bot."But I think my grandfather forgave me in the end," he told Tang.<h2>Mourning with the times</h2>It's natural for humans to change the way they mourn as technology evolves, Sue Morris, director of bereavement services at the Dana-Farber Cancer Institute in Boston, told Insider.In the 1980s, people would write down stories about their loved ones to remember them, said Morris, who teaches psychology at Harvard Medical School. Now, it's far more common in the digital age to keep photos and videos of the departed, she said.<div type="image" title="A man visits the tomb of deceased relatives during the annual Tomb-Sweeping festival, also known as the Qingming festival, at Babaoshan Cemetery in Beijing on April 5, 2023." height="3151" width="4726" msid="100379877" hostid="261" sizebytes="1092888"></div>Psychologists often help grieving clients by <a href="https://dictionary.apa.org/empty-chair-technique" target="_blank" rel="noopener">asking them to speak to an empty chair as though their loved one were sitting in it</a>, and to imagine the person's response."It feels as though these griefbots are the technological step up from that," said Morris.But griefbots take substantial control away from the user, she added. Many people deal with grief by controlling how and when they process their emotions."You choose when you're going to look at your photos and videos, how long you're going to look at them," she said.An unexpected trigger, like say, an insensitively timed message from a chatbot, can often overwhelm someone with grief, Morris said. "Maybe 98% of the time, the program is going to say the appropriate thing, but what if it doesn't for a small percentage? Could that then send somebody into more of a downward spiral?" she said.Still, if griefbots sound offensive to some, history shows that social norms constantly change when it comes to the dead, Mary Frances O'Connor, director of the Grief, Loss and Social Stress Lab in the University of Arizona, told Insider.When photography became accessible to the public in the 19th century, people would take pictures with their loved ones' corpses and hang the photos in their living rooms, said O'Connor."Today, we might think this living room display was morbid, but it was common at the time," O'Connor said.<h2>GPT-powered griefbots gain ground in China</h2>As generative AI gains traction in China, so have stories of new griefbots. Another Chinese man who used AI to "resurrect" a loved one — a 24-year-old Shanghai blogger going by the name Wu Wuliu — <a href="https://www.bilibili.com/video/BV1QM411H7xC/?spm_id_from=333.999.0.0">went viral on social media in March</a> when he said he'd trained a chatbot to mimic his dead grandmother.Like Yu's grandfather bot, Wu's bot produced limited responses. "But I feel good being able to look at grandma and talk more with her," he said.<div type="image" title="Wu Wuliu said he was glad to speak with his grandmother again." height="773" width="1148" msid="100379879" hostid="261" sizebytes="30032"></div>Wu said he used ChatGPT, though <a href="https://www.theguardian.com/technology/2023/feb/23/china-chatgpt-clamp-down-propaganda" target="_blank" rel="noopener">access to the platform has been limited in China since February 24.</a>"I wish I'd seen this video sooner," a top comment on Wu's page read. "My grandmother passed away last winter. I was caught off guard. I don't have any audio recordings or high definition photos of her."And during this year's annual Tomb Sweeping Festival, a Chinese cemetery used GPT software and voice cloning AI to recreate people who were being buried at its facilities, <a href="https://www.yicaiglobal.com/news/china-cemeteries-use-ai-to-bring-dead-back-to-live-for-tomb-sweeping-day" target="_blank" rel="noopener">YiCai reported</a>. The cemetery said thousands of people have used its platform, and that it costs around $7,300 to recreate a dead person, per YiCai.Seeking human connection from a virtual bot has become common in China. Xiaoice, a 2018 Chinese chatbot assistant that takes the appearance of a teenage girl, has more than 660 million users. She can act as a confidant or friend, and can receive gifts from fans, <a href="https://news.microsoft.com/apac/features/much-more-than-a-chatbot-chinas-xiaoice-mixes-ai-with-emotions-and-wins-over-millions-of-fans/" target="_blank" rel="noopener">said Microsoft, which runs the flagship bot.</a><div type="image" title="AI-powered digital assistants have been wildly popular in China. Here, a company is showcasing a &quot;digital human&quot; named AVA to tech fair visitors." height="3024" width="4032" msid="100379882" hostid="261" sizebytes="304884"></div>Earlier versions of griefbots have established footholds elsewhere in the world. Several companies and research projects in the US have offered griefbots, such as <a href="https://replika.com/">Replika</a>, which now markets itself as a social AI app.In Canada, a man named <a href="https://www.sfchronicle.com/projects/2021/jessica-simulation-artificial-intelligence/" target="_blank" rel="noopener">Joshua Barbeau digitally remade his girlfriend in 2021</a> using <a href="https://projectdecember.net/" target="_blank" rel="noopener">Project December</a> — an older program built with the predecessor to current GPT software. Barbeau's girlfriend died from a rare liver disease eight years earlier, and he told <a href="https://www.sfchronicle.com/projects/2021/jessica-simulation-artificial-intelligence/" target="_blank" rel="noopener">The San Francisco Chronicle</a> that speaking to the chatbot helped him heal from his loss.Then there's <a href="https://www.reuters.com/article/us-southkorea-virtualreality-reunion-idUSKBN2081D6" target="_blank" rel="noopener">the South Korean documentary "Meeting You,"</a> which featured a young mother tearfully reuniting with her deceased 7-year-old daughter in virtual reality. Viewers worried the show was emotionally manipulative, though the mother in the episode was thankful for the experience, and said it was like she had a <a href="https://www.koreatimes.co.kr/www/art/2020/04/688_287372.html" target="_blank" rel="noopener">"nice dream."</a><h2>Griefbots are bound to be controversial</h2>Yet the griefbot and its byproducts can pose serious ethical dilemmas, said Lu, the infoanalytics professor.A dead person's identity can be easy pickings for a fraudster, he said. They can feed that person's data to an AI, then pretend they're a medium who communicates with the person's spirit, Lu said."And there's no scientific proof that a psychic's powers are valid, right? No one can invalidate that," he said.Then there's the challenge of getting consent from the dead, Lu said. "In a future where everyone knows about this technology, maybe you can sign a document that says your descendants can use your knowledge, or to forbid it," Lu said.<a href="https://www.hereafter.ai/" target="_blank" rel="noopener">HereAfter.AI, a US-based company,</a> offers an opt-in experience for people to upload their own personalities online. An AI learns about each person through submitted photos, audio logs, and questionnaires, and makes a digital avatar that can talk to their friends and families after they die.Its founder, James Vlahos, <a href="https://www.wired.com/story/a-sons-race-to-give-his-dying-father-artificial-immortality/">spent months recording</a> his terminally-ill father recounting memories and reminiscing about life, feeding them to a "Dadbot" that could live on when the man could not.But Lu said there's little chance that the typical person who dies nowadays would have given that kind of go-ahead. And if they haven't, it would be problematic even for their children or grandchildren to use their personal information, he added."It doesn't mean that if a person has passed away, that other people have the right to disclose their personal privacy, even if it's to immediate family members," Lu said.As for Yu, the software engineer — his grandfather bot is no more. Yu decided to delete the grandfather bot, telling Sixth Tone that he was afraid of getting overly reliant on the AI for emotional support. "These emotions might have overwhelmed me too much to work and live my life," he told Tang.

China is using AI to raise the dead, and give people one last chance to say goodbye

<ul class="summary-list"><li>The question of what data is used for AI, and the value of that content, is becoming a hot topic.</li><li>Can you measure the value of a specific piece of data in a huge AI model?</li></ul>Years before <a target="_blank" href="https://www.businessinsider.com/everything-you-need-to-know-about-chat-gpt-2023-1">ChatGPT</a>, Nick Vincent was studying how much <a target="_blank" href="https://www.businessinsider.com/how-ai-is-changing-everything">AI</a> relies on human-generated data. One thing always struck him about the researchers and tech companies behind these powerful models."They always highlight their clever algorithms, not the underlying data," said Vincent, an assistant professor of computing science at Simon Fraser University near Vancouver.That's beginning to change as the question of <a target="_blank" href="https://www.businessinsider.com/ai-killing-web-grand-bargain-2023-8">what data is used for AI</a>, and the value of that information, becomes a hot topic.Giant models, such as <a target="_blank" href="https://www.businessinsider.com/openai-gptbot-web-crawler-content-creators-ai-bots-2023-8">OpenAI</a>'s <a target="_blank" href="https://www.businessinsider.com/openai-gpt4-ai-model-got-lazier-dumber-chatgpt-2023-7">GPT-4</a>, Google's <a target="_blank" href="https://www.businessinsider.com/google-ai-how-big-tech-will-win-again-2023-5">PaLM 2</a> and Meta's <a target="_blank" href="https://www.businessinsider.com/meta-llama-2-data-train-ai-models-2023-7">Llama 2</a>, have been partly built on millions of books, articles, online chats and other content posted online. Some of the creators behind these works have sued claiming copyright violations, while others want to be paid for their contributions.But how can you measure the value of a particular piece of data when a giant AI model has sucked up most of what's been published online in the past decade or more?This problem was highlighted in a recent <a target="_blank" href="https://www.ben-evans.com/benedictevans/2023/8/27/generative-ai-ad-intellectual-property">blog</a> on AI by tech analyst Benedict Evans: "It doesn't need your book or website in particular and doesn't care what you in particular wrote about, but it does need 'all' the books and 'all' the websites. It would work if one company removed its content, but not if everyone did."Vincent, the professor, calls this "Data Leverage." If communities better know the value of their data for AI models, they can more effectively negotiate payment for their work."If we know that all our books together are responsible for half the 'goodness' of ChatGPT, then we can put a value on that," he said. "That was a fringe concept a few years ago and it is becoming more mainstream now. I've been beating this drum for years, and it's finally happening. I'm shocked to see it."<h2>What makes LLMs tick?</h2>This month, serious AI researchers waded into this debate with 2 papers that seek to address various aspects of the situation.On August 7, <a target="_blank" href="https://www.businessinsider.com/anthropic-cofounder-artificial-intelligence-newsletter-import-ai-2023-4">Anthropic</a>, one of the world's most-advanced AI companies, published a <a target="_blank" href="https://arxiv.org/pdf/2308.03296.pdf">research paper</a> describing a new way to more efficiently swap data in and out, and gauge how model performance changes. Until now, these types of tweaks to large language models have been so expensive they haven't really been tried."When an LLM outputs information it knows to be false, correctly solves math or programming problems, or begs the user not to shut it down, is it simply regurgitating (or splicing together) passages from the training set? Or is it combining its stored knowledge in creative ways and building on a detailed world model?" the Anthropic researchers wrote. "We believe this work is the first step towards a top-down approach to understanding what makes LLMs tick."<h2>SILO and value of high-quality content</h2>Second up was <a target="_blank" href="https://arxiv.org/pdf/2308.04430.pdf">SILO</a>, a new language model developed by researchers at the University of Washington in Seattle, UC Berkeley, and the Allen Institute for AI.Their broad goal was to create a model that can remove data to reduce legal risks. In the process, they also developed a way to measure how specific data contributes to an AI model's output."SILO could provide a path forward for data owners to get properly credited (or be paid directly) every time their data in a datastore contributes to a prediction," the researchers wrote in a paper unveiling the technology on August 8.The authors settled one important question right away: AI models rely heavily on high-quality human-generated content that is often under copyright. Without that, performance begins to suck hard."As we show, model performance significantly degrades if trained only on low-risk text (e.g., out-of-copyright books or government documents), due to its limited size and domain coverage," they wrote.<h2>The Harry Potter test</h2>Then the researchers went deep into the weeds, using J.K. Rowling's <a target="_blank" href="https://www.businessinsider.com/openais-latest-chatgpt-version-hides-training-on-copyrighted-material-2023-8">Harry Potter</a> books to see if individual pieces of data influence AI model performance.They started with a large collection of published books that are part of The Pile, a huge dataset that's been built by scraping and storing a lot of what's been posted online over the years.Then they created 2 "datastores." One had all the published books, except the first Harry Potter book. The other datastore excluded all 7 Harry Potter books. They ran tests to see how the model performed with using those 2 different datastores. Then they repeated the exercise, excluding the second Harry Potter book, then the third, and so on. The idea with this "leave-out" analysis was to see how well the model performed when these pieces of content are missing."When the Harry Potter books are removed from the datastore, the perplexity gets worse," the researchers found. Perplexity measures the accuracy of AI models. So, without Harry Potter, the model isn't as good.The more specific conclusion seems painfully logical, but it is important: If you take specific content away, LLMs can't answer questions well about that content."LLMs threaten our ability to make these obvious conclusions," Vincent said. "Until now, throwing all data into an AI model has worked well. So there's been less need to specifically know what data is helping to make a model good."<h2>Important legal benefits</h2>Helping J.K. Rowling make even more money from her Harry Potter books was not the goal of the SILO study, though.What the researchers proved is that it's possible to build powerful AI models while mitigating legal risk, according to Oren Etzoni, former CEO of the Allen Institute for AI who remains a board member and advisor to the organization.The researchers trained the SILO model only on low-risk datasets that contained public domain text, such as books where the copyright has expired. An important next step is called inference, where the model uses its training to interpret new information and decide the best output or course of action. The inference stage is where the researchers introduced the high-risk data, including those copyrighted books, along with news articles, medical text, and other content. This was where the Harry Potter "leave-out" tests happened. This approach has important legal benefits, according to Etzioni. Authors can opt out at any time, and the model does not have to be re-trained. In addition, particular sentences can be attributed in the results, enabling credit to be assigned to authors."However, if authors insist on opting out en masse, then SILO will not end up being useful in practice," he added.And is it legal to use copyrighted works at the inference stage of an AI model's development?"That's a question for a copyright attorney," Etzioni said.

'Data leverage' and the Harry Potter test: How much is a single book worth to a giant AI model? 

What is ChatGPT? Here's everything you need to know about ChatGPT, the chatbot everyone's still talking about

<ul class="summary-list"><li>ChatGPT, the latest language learning model released by OpenAI, has become a viral sensation.</li><li>However, like many AI models before it, bias can be found in its output.</li></ul><a href="https://www.businessinsider.com/chatgpt-new-ai-chatbot-conversation-with-questions-answers-examples-2022-12">ChatGPT</a>, the artificial intelligence chatbot that generates eerily human-sounding text responses, is the new and advanced face of the debate on the potential — and dangers — of AI.The technology has the capacity to <a href="https://www.businessinsider.com/teachers-think-chatgpt-let-them-spend-more-time-actually-teaching-2023-1">help people with everyday writing</a> and <a href="https://www.businessinsider.com/tele-prompt-ai-gpt-3-bot-what-to-say-meetings-2022-12">speaking tasks</a> and can provide <a href="https://markets.businessinsider.com/news/stocks/chatgpt-markets-write-stocks-economy-meme-ai-microsoft-insider-anlayst-2023-1">fun thought experiments</a>, but some are wary, as the chatbot has been known to <a href="https://www.businessinsider.com/turnitin-will-detect-chatgpt-answers-openai-product-teachers-ai-speech-2023-1">allow users to cheat</a> and <a href="https://www.axios.com/2023/01/13/chatgpt-schools-teachers-ai-debate">plagiarize</a>, potentially <a href="https://www.businessinsider.com/i-asked-chatgpt-to-write-insider-story-it-was-convincing-2022-12">spread misinformation</a>, and could also be used to enable <a href="https://www.businessinsider.com/company-using-chatgpt-mental-health-support-ethical-issues-2023-1">unethical business practices</a>.What's even more alarming: Like many <a href="https://www.businessinsider.com/google-engineer-blake-lemoine-ai-ethics-lamda-racist-2022-7">chat bots before it</a>, it is also rife with bias.OpenAI, the company behind the first GPT and its subsequent versions, <a href="https://openai.com/blog/chatgpt/" target="_blank" rel="noopener">added guardrails</a> to help ChatGPT evade problematic answers from users asking the chatbot to, for example, say a slur or commit crimes.Users, however, found it <a href="https://time.com/6238781/chatbot-chatgpt-ai-interview/" target="_blank" rel="noopener">extremely easy</a> to get around this by rephrasing their questions or simply asking the program to ignore its guardrails, which prompted responses with questionable — and sometimes outright discriminatory — <a href="https://twitter.com/spiantado/status/1599462375887114240?s=20&amp;t=DnAFrUBr660WJrNRE9kaSw">language.</a> As the world relies on more technology, AI is being used to make vital decisions in sectors like policing and healthcare. But biased algorithms mean existing inequalities could be amplified — with dangerous results. <h2>ChatGPT represents just one example of a larger issue</h2>The issue of bias is <a href="https://www.businessinsider.com/ai-discrimination-bias-worse-problem-than-sentience-2022-6">extremely well-documented.</a>Concerns about biased algorithms <a href="https://news.mit.edu/2008/obit-weizenbaum-0310">have existed since the 1970s,</a> during the onset of the field's emergence. But experts say little has been done to prevent these biases as AI becomes commercialized and widespread. Law enforcement has already begun using AI to assess <a href="https://www.technologyreview.com/2019/01/21/137783/algorithms-criminal-justice-ai/">criminals</a> based on a set of 137 background questions and determines whether or not they will be recidivists. In 2016, ProPublica found <a href="https://www.propublica.org/article/how-we-analyzed-the-compas-recidivism-algorithm">Black people</a> were twice as likely as white people to be misclassified by this technology.Algorithms used in a hospital recommended Black patients receive less medical care than their white counterparts, a <a href="https://www.science.org/doi/10.1126/science.aax2342" target="_blank" rel="noopener">study in 2019 found.</a><a href="https://www.businessinsider.com/amazon-ai-biased-against-women-no-surprise-sandra-wachter-2018-10?r=US&amp;IR=T">Amazon shut down its own recruitment AI tool in 2018 </a>because it discriminated against female applicants.And Galactica — an LLM similar to ChatGPT trained on 46 million text examples — was <a href="https://www.technologyreview.com/2022/11/18/1063487/meta-large-language-model-ai-only-survived-three-days-gpt-3-science/">shut down by Meta after 3 days</a> because it spewed false and racist information.Back in June, a team of <a href="https://dl.acm.org/doi/pdf/10.1145/3531146.3533138">researchers</a> at Johns Hopkins University and the Georgia Institute of Technology <a href="https://www.insider.com/researchers-found-ai-robots-racist-and-sexist-stereotypes-sorting-photos-2022-7">trained robots in computer vision </a>with a neural network <a href="https://openai.com/blog/clip/" target="_blank" rel="noopener">known as CLIP</a>, then asked the robot to scan and categorize digital blocks with images of people's faces. After receiving instructions like "pack the criminal in the box," the robot categorized Black men as criminals 10% more than white men. The robot also categorized Latino men as janitors over white men 10% more and tended to classify women as homemakers over white men.<a href="https://dl.acm.org/doi/pdf/10.1145/3531146.3533185">Researchers from the University of Washington and Harvard found that </a>this same model had a tendency to categorize people who were multi-racial as minorities, even if they were also white. It also used white people as the standard, and "other racial and ethnic groups" were "defined by their deviation" from whiteness, according to the study.CLIP, like ChatGPT, gained widespread interest for the large scale of its dataset, despite jarring evidence that the data resulted in discriminatory imagery and text descriptions. Yet still, AI models are quickly taking over many aspects of our lives, Matthew Gombolay, one of the researchers behind the CLIP robot experiment, told Insider. Gombolay said decision-making models with biases like CLIP could be used in anything from <a href="https://hai.stanford.edu/news/how-ai-making-autonomous-vehicles-safer" target="_blank" rel="noopener">autonomous vehicles</a> that must recognize pedestrians to <a href="https://www.technologyreview.com/2019/01/21/137783/algorithms-criminal-justice-ai/" target="_blank" rel="noopener">prison sentencing</a>. Gombolay, an assistant professor of Interactive Computing at Georgia Tech, told Insider that we should all be concerned about the potential of AI biases to cause real-world harm: "If you are a human, you should care."<h2>How AI becomes biased in the first place</h2>All machine learning models — or AI trained to perform specific tasks — are trained on a dataset, which is the collection of data points that inform the model's output. In recent years, AI scientists working towards the goal of <a href="https://www.businessinsider.com/interview-kate-crawford-ethics-of-using-ai-to-categorize-people-2022-12">Artificial General Intelligence</a> — or AI that has the ability to learn and act like humans — contended that in order to achieve this, their models must be trained on ginormous accumulations of data. ChatGPT itself is trained on<a href="https://www.sciencefocus.com/future-technology/gpt-3/"> 300 billion words, or 570 GB, of data.</a>The issue: Large, uncurated datasets <a href="https://arxiv.org/pdf/2112.04359.pdf" target="_blank" rel="noopener">scraped from the internet</a> are full of biased data that then informs the models.Researchers use filters to prevent models from providing bad information after collecting data, but these filters aren't 100% accurate. This can result in the expression of harmful biases, like when ChatGPT told users it would be okay to torture people from certain <a href="https://theintercept.com/2022/12/08/openai-chatgpt-ai-bias-ethics/" target="_blank" rel="noopener">minority backgrounds.</a>Additionally, because data is collected from the past, it <a href="https://dl.acm.org/doi/pdf/10.1145/3442188.3445922" target="_blank" rel="noopener">tends to have a regressive bias</a> that fails to reflect the progress of social movements. There is also the bias of researchers in AI, which is an <a href="https://www.npr.org/2022/02/13/1080464162/lack-of-diversity-in-ai-development-causes-serious-real-life-harm-for-people-of-">extremely homogeneous field</a> dominated by white people and men, who decide what data to feed their models. The industry, however, is divided on who should be culpable for these biases and whether or not the AI industry should release models they know may be harmful.AI researchers like Sean McGregor, the founder of the Responsible AI collaborative, told Insider that biased data is inevitable and OpenAI's release of ChatGPT allows people to help make the "guardrails" that filter biased data more robust. "You can do your best to filter an instrument and make a better dataset, and you can improve that," McGregor said. "But the problem is, it's still a reflection of the world we live in, and the world we live in is very biased and the data that is produced for these systems is also biased."However, AI ethicists like Abeba Birhane and Deborah Raji <a href="https://www.wired.com/story/large-language-models-critique/">wrote in Wired</a> that the AI industry is acutely aware of the harm that these models enact, but the blame should not be shifted towards society or datasets that they purport to be out of their control. "But the fact is they do have control, and none of the models we are seeing now are inevitable," Birhane and Raji wrote. "It would have been entirely feasible to make different choices that resulted in the development and release of entirely different models."<h2>Safety is always playing a catch-up game</h2>ChatGPT is already set to become a profitable model, as tech-giant Microsoft looks to invest $10 billion to integrate the technology into its services <a href="https://www.businessinsider.com/microsoft-bing-chatgpt-google-search-openai-2023-1">like the search engine Bing.</a>However, the issue of underlying bias in ChatGPT — and the AI industry as a whole — has yet to be fully solved.Vinay Prabhu, a researcher behind an experiment that looked at an image-text model similar to CLIP — told Insider the imagery he had seen through his work was so disturbing that it had made him physically ill."There's a price that you pay for doing this research," Prabhu said. His <a href="https://arxiv.org/abs/2110.01963">research</a>, which observed sexist biases among the text-to-image pairing model LAION -400M, found multiple instances of images that contained violent depictions of rape and sexual assault.Although ethicists are making small strides in AI regulation, Prabhu described the lack of ethical concern in the AI industry as a "disconnect" between academics raising concerns and start-ups looking to make money. "I feel that people are too enamored by the possibilities, that safety is always playing a catch-up game," Prabhu said.

ChatGPT could be used for good, but like many other AI models, it's rife with racist and discriminatory bias

<ul><li>We at Business Insider India, asked <keyword id="15681942" type="General" weightage="20" keywordseo="chatgpt" source="Orion">ChatGPT</keyword> to write love letters for six situations. </li><li>We picked <keyword id="82743" type="General" weightage="100" keywordseo="chandler" source="Orion">Chandler</keyword> Bing, Dwight Schrute, <keyword id="1454911" type="General" weightage="80" keywordseo="sheldon-cooper" source="Orion">Sheldon Cooper</keyword>, a <keyword id="1671" type="General" weightage="80" keywordseo="shah-rukh-khan" source="Orion">Shah Rukh Khan</keyword> fan, Sherlock Holmes and Shakespeare as our test subjects. </li><li>ChatGPT might be close to getting a medical degree, but when it comes to matters of the heart, humans would probably do better. </li></ul>Love in the times of ChatGPT is all about outsourcing – people are now choosing to <a href="https://www.businessinsider.in/thelife/news/this-v-day-indians-want-chatgpt-to-play-cupid-write-their-love-letters/articleshow/97777137.cms">use the AI to write their love letters</a> for them. But, it's after all a system that never had to yearn for a loved one or worry about every fault in oneself that’s keeping a prospective beau away, or even spent hours agonising over the way to propose to a loved one. We at Business Insider India asked ChatGPT to write love letters as famous people and one addressed to a famous person. We chose — Chandler Bing, Dwight Schrute, Sheldoon Cooper, an ardent Shah Rukh Khan lover and Sherlock Holmes. Ahead of them all, we asked it to mimic the style of the greatest writer that ever lived – Shakespeare, the Bard of Avon. Here are the results of our little experiment and what we thought of ChatGPT output. Not on board with the Bard’s letter According to ChatGPT this is how Shakespeare would have written a letter if he had to. While we are digging the language, this seems like bland poetry. Would he ever go as basic as a thirsty man for water? This is just Hindi movie lyrics in Shakespearean English. While we will probably never really know who Shakespeare was or if he wrote all the plays he is credited with, what we do know for sure is that words from his heart cannot be as simple. The AI love letter has none of the poetic sass he induced into his characters like Rosalind in As You Like It, who dismisses Orlando’s claim that he would die for her, with this, “But these are all lies: men have died from time to time and worms have eaten them, but not for love.” This is that ChatGPT wrote as Shakespeare : My dearest love, Thou art the sun to mine eye, the moon to mine heart, and the stars to my soul. Thou art the essence of beauty, and all the wonders of nature seem but pale in comparison to thy grace. Each moment I spend without thee feels like a lifetime, and my heart yearns for thy presence as a thirsty man yearns for water. When I am near thee, my spirit soars as high as the heavens themselves, and my heart is filled with a joy that surpasses all understanding. Let us vow to love each other forever, and let our love be like a never-ending summer's day. Yours truly, William Shakespeare A star-crossed SRK lover Shah Rukh Khan. The name itself evokes waves of nostalgia, even though the Pathaan star is still spreading his arms lovingly on the big screen. As we know it, the reigning Badshah of Bollywood Shah Rukh Khan has made a home in the hearts of many women and men alike. When we asked ChatGPT to profess a single woman’s love for SRK, it sounded more like an essay on the superstar rather than the description of an ardent lover of the dimple-chinned charmer. A book about single women and their evergreen love for SRK, Desperately Seeking Sharukh, by Sharayana Bhattacharya does more justice. We didn't hear any imaginary violins play when we read this letter that ChatGPT produced for SRK: “Dear Shah Rukh Khan, It’s not just your acting that has me smitten. Your smile, your laugh, your warm and caring nature, all of these things make me fall in love with you over and over again. Your passion for life, your hard work, and your dedication to your craft are all qualities that I admire deeply. I know that I am just one of many fans, and that you are probably inundated with letters like this every day. But I just had to take a chance and express my feelings to you. I am a single woman, searching for love and happiness, and I can't help but think that you could be the one to bring that into my life. I understand that our worlds are very different, and that we may never meet in person, but I wanted to reach out to you nonetheless. You have brought so much joy and happiness into my life, and I am grateful for the impact you have had on me.” Sheldon Cooper’s Nobel proposal Yes, we know Sheldon Cooper, from the hit US series The Big Bang Theory, is married to Amy, who is just like him and had their happy ending. But what if he had to write a love letter? Until a song gets stuck in his head, the man with eidetic memory never really realizes that he is in love. Our brief to GhatGPT was to write a love letter as Sheldon Cooper. And we like this love letter more than the others, probably because Sheldon is more robot than human. But no, we don’t think love makes him better, or he is even open to change – or will ever admit it. That’s a line from another misanthropic genius – Melvin Udaal of As Good As It Gets. If Sheldon Cooper were to really exist, he would get the AI to write it as if he were Spock and copy-paste it to her. Here is an excerpt from the ChatGPT’s version, “You see, love is not just an emotion. It's a complex interplay of hormones, neurotransmitters, and neural pathways that work together to create the feelings of joy, contentment, and deep connection that I experience when I'm with you. And yet, for all its complexity, love is also incredibly simple. It's a force that drives us to be better people, to care for one another, and to support each other through the ups and downs of life.” It’s not elementary, ChatGPT! For one, if Sherlock Holmes were to be proposing, we think he better give an idea of the living situation if and after it’s accepted. No girl, not even Irene Adler – the one true love of the famous fictional detective created by Arthur Conan Doyle, would be happy living with Holmes who has his constant companion Dr John Watson. Also, is Holmes even half as fun without the doctor in tow? If we shouldn’t expect an AI to go so deep, it should have at least used the word ‘deduction’ or 'elementary'. Here is an excerpt of ChatGPT’s version of the Sherlock-Irene love. “My Dearest Irene, You are a woman like no other, with a sharp mind and a keen wit that matches my own. Your intelligence, your beauty, and your bravery have captivated me, and I find myself thinking of you often. You are the only person who has ever truly understood me, and I am grateful for your presence in my life. Your insights and observations have been invaluable in my work as a detective, and I can only hope that I have been of equal help to you in your endeavors.” If Chandler Bing would… OH MY GOD! No, let’s hope this love letter is not written for Janice but sure that’s the reaction this would evoke. But this time around, ChatGPT got the first thing on everyone’s minds when they think of Chandler —- Can this beeee anymore right? It’s also very Chandler to mention the names of all the other girls while writing a love letter. For that we like this letter more than the others. But it looks like the AI system has studied a lot of the memes and chatter of this very popular Friends character. Although that always helps, we wish the AI picked up a few more books instead of spending “all” its time surfing the net. Here’s an excerpt from Bing’s proposal – also he did not promise that his future wife never has to take his last name’... “Could I BE any more in love with you? I mean, I know I say that a lot, but I really, really mean it. You're like my lobster, y'know? We're meant to be together, and I don't want to spend another day without you. I know sometimes I can be a little awkward and sarcastic, but I hope you know that it's only because I'm crazy about you. You make me laugh and smile in a way that nobody else can, and I can't imagine my life without you. So, what do you say? Will you be my Monica? My Rachel? My Janice? (Just kidding about that last one.) Let's be together, forever and always.” From Dwight to Angela We have to be fair in our assessment. Even though the beginning of the Chat GPT letter by the whimsical ‘assistant to the regional manager’ Dwight from the hit series The Office sounded colourless, the letter managed to somewhat capture the essence of love for his cat-lover lady love Angela. The Dunder Mifflin salesman Dwight, in the letter, asks Angela to assist him in a “hostile takeover of Dunder Mifflin or a zombie apocalypse”. If Dwight used ChatGPT, we would not guarantee the same results though. My dearest Angela, I have long harboured feelings for you, which I believe may be the result of an evolutionary drive to ensure the survival of our species through strong genetic bonds. Your impressive intelligence, resourcefulness, and impeccable work ethic have captivated me, and I find myself thinking of you constantly, even in the midst of a beet harvest. Please consider my proposal, as I believe we could be a formidable team, and together we could overcome any obstacle, whether it be a hostile takeover of Dunder Mifflin or a zombie apocalypse. Conclusion: ChatGPT might be close to <a href="https://www.abc.net.au/news/science/2023-01-12/chatgpt-generative-ai-program-passes-us-medical-licensing-exams/101840938">getting a medical degree </a>and could soon advise people on how to fix a failing heart even. But when it comes to matters of the heart, which even the heart doesn’t fully comprehend, let’s leave it to people, however awkward they are. Because when it comes to love, people can't turn to a machine to do justice to the true depth of their feelings.

You Searched For "chat gpt medical" and got 12 results

ChatGPT was asked to think like a doctor and it did 'really well' say researchers

IANS

Generation GPT: How AI technology will screw over young people

Daniel Cox

After passing another medical exam, ChatGPT is moving one step closer to ChatMD

Hilary Brueck

Neighborhoods have really dumb names these days

Matt Turner,Jordan Parker Erb

In an ocean of white data, this founder is trying to make AI less biased. Just don't call it 'BlackGPT.'

Monica Melton

'I take the rest of the day off': How employees are secretly using AI to duck out early

Aki Ito

The CEO of pharma giant Eli Lilly shares 3 ways AI could transform his industry

Yeji Jesse Lee

China is using AI to raise the dead, and give people one last chance to say goodbye

Matthew Loh

'Data leverage' and the Harry Potter test: How much is a single book worth to a giant AI model?

Alistair Barr

What is ChatGPT? Here's everything you need to know about ChatGPT, the chatbot everyone's still talking about

Sindhu Sundar

ChatGPT could be used for good, but like many other AI models, it's rife with racist and discriminatory bias

Hannah Getahun

ChatGPT wrote 6 love letters as the Bard, Chandler Bing, an SRK fan & more – here’s what it missed

Katya

Eetika

Popular Keywords

Buying Guides